flopscope.numpy.float32

Single-precision floating-point number type, compatible with C ``float``.

Single-precision floating-point number type, compatible with C float.

docinfo

Character code

'f'

Canonical name

flops.single

Alias on this platform (Linux x86_64)

flops.float32: 32-bit-precision floating-point number type: sign bit, 8 bits exponent, 23 bits mantissa.